Show HN: Vibe Coded Text Categorizer (github.com)

The Vibe Coded Text Categorizer, a new local-first CLI tool, has been launched to facilitate efficient and private text categorization through vector embeddings. This utility integrates a local embedding provider, such as LM Studio, with SQLite and sqlite-vec for rapid semantic categorization without the need for cloud processing. Its significance lies in empowering developers to manage sensitive data, such as banking transactions and logs, on their local systems, ensuring data privacy and security. Key features include high-performance vector similarity searches, simple Unix-like commands, and a portable data format stored in a single SQLite file for easy backup or migration. The tool supports importing and exporting data in JSONL and CSV formats, enabling straightforward data migration and automated labeling. Users can categorize text strings, teach the system with labeled examples, and manage their databases comprehensively. Designed by Gemini in collaboration with Tadas Vilkeliskis, this tool highlights advancements in privacy-focused AI applications, catering to those who prioritize local data processing in their development workflows.